信息与智能2023NO.8SCIENCE&TECHNOLOGYINFORMATION科技资讯SCIENCE&TECHNOLOGYINFORMATION科技资讯多个经验模态分解对振动信号作用的对比赵楠石振刚*(沈阳理工大学信息科学与工程学院辽宁沈阳110158)摘要:经振动传感器采集到的信号是非线性、非稳定的,这种信号无论是在时域还是频域上都不易分析。所以通过经验模态分解将原始信号分解成为多个本质模态函数(IntrinsicModeFunction,IMF),之后对其进行特征提取等进一步处理。但是经验模态分解存在模态混叠与端点效应的问题,所以文章采用互补集合经验模态分解(ComplementaryEnsembleEmpiricalModeDecomposision,CEEMD)。CEEMD是在进行经验模态分解之前加入多组符号相反的白噪声,这不仅减少了模态混叠,分解出的IMF分量还更精进。这种互补集合经验模态分解有效地处理了所采集的非线性、非稳定性的振动信号。关键词:振动信号经验模态分解本质模态函数互补集合经验模态分解中图分类号:TN911文献标识码:A文章编号:1672-3791(2023)08-0017-04ComparisonoftheEffectsofMultipleEmpiricalModeDecompositiononVibrationSignalsZHAONanSHIZhengang*(SchoolofInformationScienceandEngineering,ShenyangLigongUniversity,Shenyang,LiaoningProvince,110158China)Abstract:Thesignalcollectedbyvibrationsensorsisnonlinearandunstable,andthiskindofsignalisdifficulttoanalyzeineithertimedomainorfrequencydomain.Therefore,theoriginalsignalisdecomposedintomultipleIMFsthroughempiricalmodedecompositionandthenisfurthercarriedoutprocessingsuchasfeatureextraction.How‐ever,modalaliasingandendpointeffectsexistinempiricalmodedecomposition,sothispaperadoptsCEEMD.CEEMDistoaddmultiplegroupsofwhitenoiseswithoppositesignsbeforeempiricalmodedecomposition,whichnotonlyreducesmodealiasing,butalsomakesthedecomposedIMFcomponentmorerefined.Thecomplementaryensembleempiricalmodedecompositioneffectivelydealswiththecollectednonlinearandunstablevibrationsignals.KeyWords:Vibrationsignal;Empiricalmodedecomposition;Intrinsicmodefunction;Complementaryensembleempiricalmodedecomposition经振动传感器采集得到的振动信号,是一个非线性、非平稳的信号。通常于信号的时域或频域上对信号进行分析与特征提取。而振动信号在时域上无法观测出信号的特征,在频域上,能看...